An application of Artificial Bee Colony algorithm and comparison with its variants

Artificial Bee Colony (ABC) is a simple and robust algorithm in finding optimal solutions to numerous optimization problems. ABC algorithm performs better than other population based single objective optimization algorithms and it requires only few control parameters. ABC algorithm has been extensively used in applications such as continuous optimization, structural optimization, combinatorial optimization and many more. In this research, ABC algorithm has been applied for the first time to reconstruct the gene regulatory network from gene expression data. Researchers have incorporated mutation and crossover operator with the original ABC algorithm to enhance its performance. Experiment results shows that the original ABC algorithm outperforms some of its variants in inferring genetic network underlying microarray data. Linear time variant model has been used to reverse engineer the gene regulatory network. The proposed approach has been tested on noise free time series datasets. Then, it was tested on noisy time series datasets. The proposed reconstruction technique has been further validated by analyzing the SOS DNA repair network in Escherichia coli. The proposed ABC based inference method have shown its strength in discovering reasonable regulations compared to some of the previous research in inference of the genetic networks.